[ISSUE-154][Improvement] Support Empty assignment to Shuffle Server by rhh777 · Pull Request #325 · apache/uniffle · GitHub
Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@

package org.apache.spark.shuffle;

import java.util.Collections;
import java.util.List;
import java.util.Map;
import java.util.Set;
Expand Down Expand Up @@ -269,6 +270,17 @@ public <K, V, C> ShuffleHandle registerShuffle(int shuffleId, ShuffleDependency<
}
LOG.info("Generate application id used in rss: " + id.get());

if (dependency.partitioner().numPartitions() == 0) {
LOG.info("RegisterShuffle with ShuffleId[" + shuffleId + "], partitionNum is 0, "
+ "return the empty RssShuffleHandle directly");
return new RssShuffleHandle(shuffleId,
id.get(),
dependency.rdd().getNumPartitions(),
dependency,
Collections.emptyMap(),
RemoteStorageInfo.EMPTY_REMOTE_STORAGE);
}

String storageType = sparkConf.get(RssSparkConfig.RSS_STORAGE_TYPE.key());
RemoteStorageInfo defaultRemoteStorage = new RemoteStorageInfo(
sparkConf.get(RssSparkConfig.RSS_REMOTE_STORAGE_PATH.key(), ""));
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -17,18 +17,34 @@

package org.apache.uniffle.test;

import java.util.ArrayList;
import java.util.Iterator;

import com.google.common.collect.Lists;
import org.apache.spark.api.java.JavaPairRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.Function;
import org.apache.spark.api.java.function.Function2;
import org.apache.spark.api.java.function.PairFlatMapFunction;
import scala.Tuple2;

public class TestUtils {
private TestUtils() {

}

static JavaPairRDD<String, Integer> getEmptyRDD(JavaSparkContext jsc) {
JavaPairRDD<String, Integer> javaPairRDD1 = jsc.emptyRDD().flatMapToPair(
new PairFlatMapFunction<Object, String, Integer>() {
@Override
public Iterator<Tuple2<String, Integer>> call(Object s) throws Exception {
return new ArrayList<Tuple2<String, Integer>>().iterator();
}
}
);
return javaPairRDD1;
}

static JavaPairRDD<String, Integer> getRDD(JavaSparkContext jsc) {
JavaPairRDD<String, Integer> javaPairRDD1 = jsc.parallelizePairs(Lists.newArrayList(
new Tuple2<>("cat1", 11), new Tuple2<>("dog", 22),
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@

import java.io.OutputStreamWriter;
import java.io.PrintWriter;
import java.util.Collections;
import java.util.concurrent.TimeUnit;

import com.google.common.util.concurrent.Uninterruptibles;
Expand Down Expand Up @@ -97,6 +98,16 @@ public void test() throws Exception {
assertEquals(remoteStorage1, remoteStorageInfo1.getPath());
assertTrue(remoteStorageInfo1.getConfItems().isEmpty());

// emptyRDD case
JavaPairRDD<String, Tuple2<Integer, Integer>> javaEmptyPairRDD1 = TestUtils.combineByKeyRDD(
TestUtils.getEmptyRDD(jsc1));
ShuffleDependency emptyShuffleDependency1 = (ShuffleDependency) javaEmptyPairRDD1.rdd().dependencies().head();
RssShuffleHandle emptyRssShuffleHandle1 = (RssShuffleHandle) emptyShuffleDependency1.shuffleHandle();
assertEquals(javaEmptyPairRDD1.rdd().dependencies().head().rdd().getNumPartitions(), 0);
assertEquals(emptyRssShuffleHandle1.getPartitionToServers(), Collections.emptyMap());
assertEquals(emptyRssShuffleHandle1.getRemoteStorage(),RemoteStorageInfo.EMPTY_REMOTE_STORAGE);


// the same app would get the same storage info
JavaPairRDD<String, Tuple2<Integer, Integer>> javaPairRDD2 = TestUtils.combineByKeyRDD(TestUtils.getRDD(jsc1));
ShuffleDependency shuffleDependency2 = (ShuffleDependency) javaPairRDD2.rdd().dependencies().head();
Expand Down